This article focuses on the advantages of organizing rules for knowledge extraction in a hierarchical order. It details the methodology to develop extraction rules, and illustrates the results with examples taken from real-life competitive intelligence applications.